Azoospermia, defined as the absence of sperm cells in the ejaculate, is a phenomenon detected in about 1% of the male population and 10 – 15% of infertile men [1, 2]. Unlike obstructive azoospermia, in which there is an obstruction of the spermoducts, non-obstructive azoospermia is characterized by a complete absence of sperm cells in the seminal fluid due to minimal spermatogenesis or its absence as well. Men suffering from non-obstructive azoospermia represent the most complicated infertile men category to treat. Compared to fertile men, they usually have an increased level of follicle-stimulating hormone (FSH), a reduced level of total testosterone in the blood serum, and a reduced testicular volume. So, to initiate a biological pregnancy, they usually require in vitro fertilization (IVF) and intracytoplasmic sperm injection (ICSI) through using testicular sperm which has been obtained surgically [3]. Unfortunately, the frequency of surgical sperm extraction in men with non-obstructive azoospermia remains low [4].

Histopathological diagnoses based on testicular biopsy in men with non-obstructive azoospermia include Sertoli cell-only syndrome (SCO-syndrome), hypospermatogenesis, and the sperm maturation arrest, determined by the presence of germ cells that have not reached full maturity. A homogeneous stoppage of sperm maturation is characterized by a stoppage of spermatogenesis at the same stage in all the seminal tubules [5]. It is divided into an early one (so that only spermatogonia or spermatocytes are detected) and a late one (spermatids without spermatozoa are detected) [6][7]. Some men with sperm maturation arrest may have spermatogenesis foci in the testicles, and the spermatozoa found in these loci during microTESE can be used for IVF/ICSI [5].

The sperm maturation arrest can be a primary (genetic or idiopathic) or acquired one. Thus, acquired causes include iatrogenic conditions (chemotherapy, radiation therapy, medication therapy, and testosterone supplements), past infections, endocrinopathies, congenital anomalies (cryptorchidism), testicular torsion, and varicocele [8]. Such patients, as well as patients suffering from obstructive azoospermia, have normal serum hormone levels (FSH, luteinizing hormone, testosterone, and prolactin) and testicular volume [9]. But there are also contradictory data. So, T. Ishikawa et al. It was found that individuals with later stages of sperm maturation arrest had a lower level of FSH and a larger diameter of the seminal tubules than men with earlier stages [8]. It has also been reported that genetic abnormalities, such as Y-chromosome microdeletions and karyotypic anomalies, are more often detected in patients with sperm maturation arrest [5][10][11][12].

Therefore, Hung et al. reported that men with uniform sperm maturation arrest and normal FSH had a lower frequency of surgical sperm extraction using microTESE and worse IVF/ICSI results than other men suffering from non-obstructive azoospermia [5]. For men with obstructive azoospermia, for example, those who have previously undergone vasectomy, the frequency of surgical sperm extraction should be almost 100% [13]. However, even though nomogram predictions depending on testicular size and serum FSH levels are sometimes effective for differentiating men with non-obstructive azoospermia and idiopathic obstructive azoospermia, they are not entirely accurate [14]. In addition, men with the histology of sperm maturation arrest may have normal testicular volume and a relatively lower level of FSH compared to other histological subtypes of non-obstructive azoospermia, which may complicate the differential diagnosis with obstructive azoospermia [15].

Thus, a marker that allows differentiating these two conditions (obstructive azoospermia and sperm maturation arrest) would help in predicting the success of surgical methods of sperm extraction and counseling patients with male infertility. A potential source of such biomarkers is seminal plasma.

Therefore, this study aimed to evaluate the lipidomic profile of ejaculate in patients suffering from sperm maturation arrest.

The authors of this article studied the lipidomic profile of seminal plasma of 64 patients with azoospermia and 24 healthy men with normozoospermia (as a control group). This study was carried out at the Andrology and Urology Divison, Kulakov National Research Medical Center of Obstetrics, Gynecology, and Perinatology in 2019 – 2021. This study was also approved by the Ethics Committee of Sechenov University. All the patients gave written consent to participate in the study.

The study group included men over the age of 18 with azoospermia, confirmed twice by the seminal fluid analysis. Exclusion criteria: infectious diseases (acute and chronic in the acute stage), retrograde ejaculation, and anejaculation. All 64 men underwent the microTESE procedure, which was performed following standard protocols described by Dabaja and Schlegel [16], followed by a pathomorphological examination of testicular biopsies. In this study, the seminal plasma lipid profile of men with non-obstructive azoospermia, including histology of sperm maturation arrest, was evaluated.

Lipids from the seminal plasma were isolated by Folch extraction. Lipid extracts and quality control samples were analyzed on liquid chromatography according to the method described earlier by the authors [17]. Lipids were identified by using the Lipid Match R-script1 by exact mass using the Lipid Maps database2 and by characteristic tandem mass spectra (MS/MS).

Statistical analysis. As for statistical processing of the results, the authors used scripts written in the R language3 and the RStudio program4 (RStudio PBC, USA). Before the study, the data were normalized to the median values of the corresponding peaks in the quality control samples. While comparing the “control” and “azoospermia with sperm maturation arrest” groups, the Mann-Whitney test was used. The Wilcoxon criterion was used to compare the lipid profiles of the ejaculate before and after surgery. Median (Me) and quartiles Q1 and Q3 were used to describe quantitative data. The value of the threshold significance level p was assumed to be 0.05.

The selection of variables for the construction of diagnostic models based on the logistic regression “control” / “stopping sperm maturation” was carried out by a two-stage method: by using discriminant analysis, orthogonal projections of variables on hidden structures (OPLS-DA) determined the values of projections of variables (PV) and selected those that satisfied the condition PV > 1. Variables were selected from them step by step, based on the value of the Akaike information criterion (ICA). When the growth of ICA stopped, those whose coefficients did not differ statistically significantly from 0 (the significance threshold was 0.05) were excluded from the selected set of compounds step by step. The obtained models were validated by using cross-validation for a separate object.

Therefore, after microdissection testicular biopsy, sperm cells were found in 25 out of 64 patients. Thus, 14 out of 64 patients had a pathomorphological picture of sperm maturation arrest and all had a negative outcome of microTESE.

While comparing the control group and the group with sperm maturation arrest, 22 lipids in the positive ion mode were selected as statistically significant ones (Table 1) and 11 lipids in the negative ion mode (Table 2). Lipids belong mainly to the classes of hexosylceramides, sphingomyelins and phosphatidylcholines – esters, and oxidized lipids.

Diagnostic models based on logistic regression were constructed as well to determine the stop of sperm maturation in the mode of positive and negative ions (Tables 3-4, Fig. 1).

Thus, according to the results of cross-validation, the model in the positive-ions mode had more favorable prognostic characteristics, namely good sensitivity (91%) and satisfactory specificity (85%). In the negative-ion mode, according to the results of cross-validation, the model had less favorable prognostic characteristics, namely satisfactory sensitivity (75%) and specificity (81%).

While consulting about infertility for patients with non-obstructive azoospermia, it is important to provide information about the chances of sperm-obtaining. An unsuccessful microTESE procedure, especially with simultaneous oocyte retrieval, can have irreparable emotional and financial consequences for both members of a married couple [18]. According to the earlier studies, the use of histopathological patterns can be considered as a valuable predictor of sperm extraction [15, 19, 20].

In this study, the authors evaluated the lipid profile of ejaculate in patients with sperm maturation arrest. The authors were able to find candidate lipids, including PC 16:0_22:6, PC O-16:1/16:0, SM d20:1/22:2, and PG 20:3_22:6, belonging to the classes of phosphatidylcholines, sphingomyelins (phospholipids), and phosphatidylglycerines, which in multivariate analysis turned out to be statistically significant predictors of non-obstructive azoospermia with sperm maturation arrest. Sicchieri et al. found that when cryopreserved spermatozoa were thawed, their overall motility significantly increased after treatment with L-phosphatidylcholine and L-acetyl-carnitine [21]. Also, Vireque et al. demonstrated in their study that Lα-phosphatidylcholine improves the quality of sperm cells in vitro [22]. Boguenet et al. analyzed the metabolomic profile of seminal plasma of 20 men suffering from severe oligoastenozoospermia and compared it with the one of 20 men with normozoospermia. Therefore, they revealed a decrease in concentrations of 17 phosphatidylcholines and 4 sphingomyelins in the group with severe oligoastenozoospermia [23].

Sphingomyelins are components of prostasomes that, being fused with sperm cells, stabilize their plasma membrane, enriching it with cholesterol, sphingomyelin, and saturated glycerophospholipid. This prevents the premature occurrence of an acrosomal reaction [24]. Rivera-Egea et al. analyzed the lipid composition of sperm cells from infertile patients after intracytoplasmic sperm cells injection (the group of non-pregnant (n = 16) and compared it with the group of pregnant (n = 22)). As a result, 151 different lipids were found in samples, 10 of which were significantly increased in samples from the group of non-pregnant, ranging from 1.10 to 1.30 times. Primarily, these were ceramides, sphingomyelins, and three glycerophospholipids, one lysophosphatidylcholine, and two types of plasmalogens [25].

Discussing the faults of the study, the authors need to touch upon the choice of the control group. Normozoospermia will differ from any azoospermia in the metabolomic profile due to the absence of sperm cells. And it is the disadvantage of this study. The authors decided that they could not use obstructive azoospermia as a control group, since the lipidome would depend on the level of obstruction. Considering the heterogeneity of non-obstructive azoospermia, this group is also not suitable. It is also doubtful to use separate subspecies of non-obstructive azoospermia as a comparison group since their metabolomic profile has not been described in detail before. Anyway, this study is the first one of such a series. The authors decided to use normozoospermia as a control group. The distortion of the results due to the influence of mature sperm cells was leveled by examining the seminal plasma separated from them. In the future, the authors also plan to compare the metabolomic profile of seminal plasma of non-obstructive azoospermia subspecies.

Even though the early stages of spermatogenesis are equally preserved in both fertile men and men with homogeneous sperm maturation arrest, the ejaculate in the studied group of patients differs according to its lipid profile. Patients suffering from non-obstructive azoospermia, against the background of meiosis arrest, may have unique lipidomic characteristics of seminal plasma, which may in the future allow differentiating various variants of severe male infertility by using non-invasive methods.

1. Koelmel JP, Kroeger NM, Ulmer CZ, Bowden JA, Patterson RE, Cochran JA, Beecher CWW, Garrett TJ, Yost RA. LipidMatch: an automated workflow for rule-based lipid identification using untargeted high-resolution tandem mass spectrometry data. BMC Bioinformatics. 2017;18(1):331. DOI: 10.1186/s12859-017-1744-3 2. Sud M, Fahy E, Cotter D, Brown A, Dennis EA, Glass CK, Merrill AH Jr, Murphy RC, Raetz CR, Russell DW, Subramaniam S. LMSD: LIPID MAPS structure database. Nucleic Acids Res. 2007;35(Database issue):D527-32. DOI: 10.1093/nar/gkl838 3. Team, R.C. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Available online: https://www.r-project.org/ 4. Team, Rs. RStudio: Integrated Development for R. Available online: http://www.rstudio.com/